Verbal protocols are widely used in user studies for evaluating websites. This study investigated the effectiveness and efficiency of concurrent and retrospective verbal protocols (CVP and RVP) for both blind and sighted participants, as well as participant workload and attitudes towards these methods. Eight blind and eight sighted participants undertook both protocols in a website evaluation. RVP was more effective as measured by problems encountered for both groups, although it was no more efficient than CVP. The severity of problems identified by both protocols was equivalent. As measured on the NASA TLX, participants found RVP found more demanding than CVP. Sighted participants found rating problems during CVP more disruptive than blind participants. These results show that RVP is a more useful protocol for practitioners and researchers even though it takes more time and is more demanding for participants. It is equally applicable for both blind and sighted participants.